We are using mutant DNA polymerases obtained by recombinant DNA technology to examine the relationship between the structural and kinetic properties of DNA polymerases and their processivity and fidelity. Emphasis is on enzymes for which structural information is available, including Klenow polymerase and DNA polymerase beta. We have determined the processivity and error specificity of exonuclease-deficient Klenow polymerase and 22 mutant derivatives altered in several key parameters. Several of these mutants have altered processivity. Two have strongly reduced fidelity, one of which selectively enhances misincorporation of pyrimidine dNTPs. Another has strongly increased fidelity. We are now determining if this reflects enhanced discrimination during dNTP insertion and/or mispair extension or during binding to paired or mispaired template-primers. The Klenow data suggest that dNTP and metal binding residues are important for determining substitution fidelity, but not template-primer misalignment-initiated errors. We have also identified a Klenow frameshift mutator polymerase with reduced processivity that results from deletion of amino acids in the polymerase "thumb" domain. We have determined the effects on fidelity of a polymerase accessory protein that enhances processivity, demonstrating a strong role in controlling frameshift fidelity. We have establish the fidelity of the wild-type and exonuclease-deficient forms of several polymerases to examine the relationship between exonucleolytic proofreading and error rates during copying of simple repeat sequences. This study shows that the contribution of proofreading to frameshift fidelity in homopolymeric runs decreases as the run length increases. The study is now being extended to the di- and tri-nucleotide repeats that are unstable in cancer cells and several hereditary diseases. We have initiated a study of mutant forms of beta polymerase based on the only structural information available for a ternary enzyme[unreadable]DNA[unreadable]dNTP complex. Emphasis is currently on amino acid residues hypothesized to make contacts important for nucleotide selectivity. It is our belief that structure function studies of DNA polymerases will improve our understanding of how the human genome is stably replicated and maintained, and how DNA adducts affect gemone stability.